<jarlg>
I have a git project where I often have to change branches and rebuild, and some of the rebuilds are unnecessary (e.g. switching back and forth between branches with no edits). With Nix, would it be possible to cache these builds? I know it's possible to hack something together by tagging builds with their "git write-tree" hash, but I'm hoping Nix can get me this?
<cole-h>
If literally nothing changed, yes, Nix won't rebuild. However, there are some non-obvious things that can change, depending on how you fetch your source.
<jarlg>
Some things may change -- is there a way to tell Nix to use "git write-tree" for example? That gets you a hash of the git index, which is what matters.
<jarlg>
Also I'm not sure what you mean by fetching sources, this is a local git repo.
<noonien>
use it how? nix doesn't really care about git, it just cares about dependencies.
<cole-h>
To build something with nix, you must specify a source. There are various ways to consume this source (e.g. fetchgit, which will use the underlying .git pristine for the files). However, since .git is stateful, there's a chance that using `fetchgit ./.` will end up pulling in "something that changed", signalling to Nix the project needs to be rebuilt.
<noonien>
what are you trying to cache? the nix configuration, or the build of your program?
<jarlg>
noonien: The build of the program.
<noonien>
oh, i see, and you're saying that even though none of the source code changes, you still get rebuilds?
rb2k has joined #nixos
<noonien>
well, you go to a different branch, so the code changes, but you say you've already had this built, and it builds again?
<jarlg>
Currently I'm not using Nix, I'm looking into Nix to see if I can avoid exactly those rebuilds you mention :)
<jarlg>
cole-h: I had "src = ~/git/project" in mind, I'm not sure how else to specify a local source.
<tejing>
yeah, I think it should avoid that. if you're building the same commit id, nix should be able to cache it correctly
<noonien>
well, nix produces builds that have a hash of all the dependencies used to build that package. as long as you do one build, and the dependencies don't change (your source being a dependency), there will not be another build
<jarlg>
tejing: That's what I was hoping!
<noonien>
one thing that might fail is if you put your .git repo into the source code
<cole-h>
jarlg: `src = ~/git/project` will pull in that entire directory, including the .git (which, is stateful)
<noonien>
the .git directory will probably contain a log of you switching branches, which means the source is different, which will trigger a different build
<noonien>
s/fail/trigger a build
<jarlg>
Yeah, I understand. Can I use fetchgit for a local source? Or what should I do?
<cole-h>
Yes, you can use `fetchgit ./.`
<jarlg>
Ah, great :) I'll try this out then, thanks!
<tejing>
if nothing else, if you use the git repo as a flake input you shouldn't have any of that trouble
<noonien>
i'm not suggesting to do this if fetchgit works for you, but you can also filter the source if you'll ever need to
<tejing>
it will probably work without flakes too though
bitmapper has joined #nixos
<jarlg>
Alright, good to know. Will try fetchgit first though.
<infinisil>
jarlg: cole-h: It's fetchGit though, with a capital G
<infinisil>
Well the builtin one at least, builtins.fetchGit, or fetchGit (same thing)
<cole-h>
Yeah, I meant `builtins.fetchGit`
<cole-h>
`pkgs.fetchgit` requires a hash
<jarlg>
Alright. Another thing: I'd like to bump the version in nixpkgs for a package I'm using. However I don't have hardware/time to run all the tests. Should I still submit a PR?
<{^_^}>
[nixpkgs] @petabyteboy pushed to master « redis: make masterAuth option optional (#109417) »: https://git.io/JtUvH
gustavderdrache has quit [Quit: Leaving.]
ZaraChimera has joined #nixos
pjt_tmp has joined #nixos
rnea has quit [Ping timeout: 260 seconds]
<cole-h>
jarlg: Go for it. ofborg (our CI) will automatically build it, at least on x86 and aarch64 linuxes
<cole-h>
jarlg: Please read the CONTRIBUTING document before you submit it though; it'll make both your and our lives easier!
<{^_^}>
[nixpkgs] @petabyteboy merged pull request #109418 → nixos/tests/docker: wait for docker service → https://git.io/JtJjC
<{^_^}>
[nixpkgs] @petabyteboy pushed to master « nixos/tests/docker: wait for docker service (#109418) »: https://git.io/JtUJT
<jarlg>
cole-h: Thanks, reading as we speak (write?) :)
<noonien>
hmm, i'm trying to get krops to work with my flaked nixos config and i'm pretty sure this is a nix error, but i don't know how to start debugging: `access to path '/etc/hostname' is forbidden in restricted mode` any tips?
<noonien>
yes, i saw that, but it just tries reading it, it doesn't throw that error
ericsagnes has quit [Ping timeout: 260 seconds]
pitch has quit [Ping timeout: 264 seconds]
<cole-h>
The error is thrown by nix
Acou_Bass has joined #nixos
<cole-h>
Because flakes don't allow accessing files outside the flake repo
<noonien>
i'm guessing it's probably because that's called on the machine that does the deploying
pitch has joined #nixos
<noonien>
i think that's it. and because i'm calling krops.lib.mkTarget, which seems to want to read the hostname
<cole-h>
You could make your own `mkTarget` using the logic present in krops, but preventing the reading of /etc/hostname and other impurities (like env vars)
<noonien>
yeah, turns out just providing the username works just fine
<noonien>
kind of late to discover sudo = true; still complains that it requires a terminal
<bqv>
adisbladis: gonna switch back to fish >_> but not because xonsh isn't good, just because there's an extremely irritating python bug that the dev refuses to solve, and i almost exclusively use emacs as my "shell" at this point anyway, so there's not much use overengineering my fallback shell
<cole-h>
What exactly are you looking for? hiroshi[m]
<pbb>
noonien: did you think about building the system configuration on the 'source' host (where you run the deploy script)?
<hiroshi[m]>
cole-h: sway setting is fine. but I don't know if sway in nixos will install a login manager. but I will read the link that you pasted. thanks
<hazel[m]>
i just put `exec sway` in my zprofile (i use zsh).
<cole-h>
Sway does not install (nor does it recommend) a login manager.
<hazel[m]>
although i use gnome now.
<cole-h>
You're expected to login on a TTY and `exec sway`
<hazel[m]>
* although i use gnome+gdm now.
<hiroshi[m]>
awesome.that is perfect
<noonien>
pbb: that's exactly what i want to avoid, and why i can't use deploy-rs, or nixus. i do think i might be able to make krops work, if i can make the ssh command spawn a terminal
Jd007 has joined #nixos
<pbb>
okay i see
<pbb>
and do you want to evaluate it on the source host?
<noonien>
i'm triggering the deploys from a meager laptop
<pbb>
so the source host is not even running nix? or is it too slow?
<noonien>
well, ideally yes, but that might be harder to do than it's worth
<cole-h>
noonien: Why not use nixos-rebuild --target-host?
<noonien>
cole-h: --use-remote-sudo requires a terminal, which nixos-rebuild doesn't spawn when executing ssh
<cole-h>
Have you tried NIX_SSHOPTS to make it spawn a tty?
<noonien>
hmm, i have not
<pbb>
because of these limitations I ended up editing my nixfiles on a powerful remote host via ssh, and deploying from there
slack1256 has quit [Remote host closed the connection]
<pbb>
basically, when you add the assumption that your source host is powerful, everything gets much easier
<{^_^}>
[nixpkgs] @veprbl opened pull request #109422 → binutils: enable gold when cross compiling on darwin → https://git.io/JtUUS
<cole-h>
Which means I don't need to specify --build-host!
Acou_Bas- has joined #nixos
Acou_Bass has quit [Ping timeout: 240 seconds]
<noonien>
i'm still working on making my nixos-config a bit more complete, some machines i build remotely, some locally, some i need to cross-compile
<noonien>
all pretty easy to do with scripts
<pbb>
noonien: no problem. when this module is included, you can run `nix build -f "<nixpkgs/nixos>" config.system.build.deployScript -I nixos-config=$host/configuration.nix && ./result switch`. Or as I did create a small redirect in a default.nix, to make it `nix build -f . deploy.$host && ./result switch` from my nixfiles directory
<noonien>
oh, my config is already flaked
<pbb>
should be very similar with flakes
<noonien>
so a `nix run ".#deploy.host"` would do the trick
<pbb>
yes
<noonien>
i create the .deploy.host app myself
<noonien>
hmm, how does nixos-rebuild --target-host work?
<pbb>
hehe
<pbb>
it depends
<pbb>
if your build-host is the same
<noonien>
yes, by default that's the case
<pbb>
you can set target-host and build-host independenly
<noonien>
i'm curious how the configuration is copied to the target host
<pbb>
if it's the same, it will evaluate the config on your source host
<pbb>
so it will build .drv files
<pbb>
copy those drv files to the target host
<pbb>
and realise (=build) them there
<noonien>
and then it just realises the .drv?
<noonien>
i see
<pbb>
that's why I asked if you want to evaluate on the source host
<noonien>
i thought it would be more involved, wanted to know of an easy solution first :D
<pbb>
I can recommend reading the source of nixos-rebuild
<pbb>
`vim $(which nixos-rebuild)` or something
<pbb>
things are much easier when you understand what's going on
zebrag has quit [Quit: Konversation terminated!]
<pbb>
now why I really opened my irc client: does anyone know why the nixos.tests.networking.scripted.virtual job is failing in hydra? I can not reproduce the failure locally
<pbb>
it should be the last remaining failing job for nixos-unstable to advance
<cole-h>
It's very flaky (no pun intended)
<cole-h>
A lot of the networking tests are
<pbb>
meh
<cole-h>
And it sucks because it frequently blocks channels
<pbb>
yeah
<pbb>
I guess it behaves just like networking on real machines :D
<cole-h>
lol
<pbb>
and also I just saw now some more tests are failing again :(
<noonien>
as far as i can tell, the derivation is copied to the target host with `nix-copy-closure --to host .#nixosConfigurations.host` in my case
<pbb>
oh?
<pbb>
so build-host is localhost?
<pbb>
or is .#nixosConfigurations.host the .drv somehow?
<noonien>
no, i mean the derivation of the configuration
<noonien>
yup
<pbb>
okay. weird flake syntax or did you write it that way in the chat as pseudocode?
<noonien>
nope, just looked at nixos-rebuild. `nixosConfigurations` is a flake output
<noonien>
the .host is actually .<host>
<noonien>
which contains the host name
<noonien>
in my case, both .<host> and --to <host> match
<noonien>
oh, actually, no, i'm wrong
zebrag has joined #nixos
<noonien>
i guess nix-copy-closure expects a nix store path
<pbb>
For non-flake configs it runs nix-instantiate in line 206. not sure if the same code is used for flakes as well? But nix-instantiate probably can't evaluate flakes?
<{^_^}>
[nix-mode] @matthewbauer merged pull request #119 → Make nix-drv-mode a derived-mode of javascript-mode → https://git.io/JLSwd
<{^_^}>
[nix-mode] @matthewbauer pushed 3 commits to master: https://git.io/JtUTa
<pbb>
only place where $flakeAttr is used as far as i can tell is l. 436
<{^_^}>
[nix-mode] @matthewbauer pushed 2 commits to master: https://git.io/JtUTw
rb2k has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<{^_^}>
[nix-mode] @matthewbauer merged pull request #120 → Use the located nixfmt binary for formatting. → https://git.io/JLbTV
<{^_^}>
[nix-mode] @matthewbauer pushed 2 commits to master: https://git.io/JtUTo
<{^_^}>
[nix-mode] @matthewbauer merged pull request #122 → Add version to invocation of define-obsolete-function-alias → https://git.io/JLNR7
<{^_^}>
[nix-mode] @matthewbauer pushed 2 commits to master: https://git.io/JtUT6
<pbb>
noonien: I haven't read nixos-rebuild since the --flake option got added. I would love to know how it evaluates the flake (without building it) when the build-host is not localhost, I couldn't figure it out for now.
<noonien>
i have a different version than yours probably, i have a few more instances of `flake` in my nixos-rebuild
<colemickens>
that function is my heavy lifter. It builds flakey nixos configs on all sorts of remote builds.
<colemickens>
every remote command is invoked through nix-shell, so if it works once, it should work reliably, even with older nixos builders
<colemickens>
it ... handles lots of edge cases, I've reimplemented it a few times, every time thinking it was overly complicated and yet it winds up looking similar each time (some of that's bash's fault)
ZaraChimera has joined #nixos
Supersonic112 has joined #nixos
Supersonic has quit [Ping timeout: 258 seconds]
Supersonic112 is now known as Supersonic
<{^_^}>
[nixos-homepage] @garbas pushed 0 commits to nixcon-fix: https://git.io/JtUIO
rb2k has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
heyitsrama has quit [Remote host closed the connection]
ZaraChimera has quit [Quit: Connection closed]
xelxebar has quit [Remote host closed the connection]
<superbaloo>
if a configuration imports from nixpkgs/nixos/modules/profiles/base.nix, and base.nix sets "zfs" in boot.supportedFilesystems. Is there a way to forcibly remove zfs from the array
<superbaloo>
it looks like everything is merged together
<gchristensen>
lib.mkForce []; will make the arra yempty
gustavderdrache has quit [Quit: Leaving.]
<superbaloo>
ho right
xelxebar has joined #nixos
<superbaloo>
yeah
<superbaloo>
that works as expected <3
<DigitalKiwi>
why you don't want zfs :(
<superbaloo>
on a raspberry pi?
domogled1 has joined #nixos
<superbaloo>
they dont exactly have proper block devices :x
<gchristensen>
I do it, but it is a bit of a stretch :)
<superbaloo>
not use it's worth it, given how ram ungry zfs is :x
domogled has quit [Ping timeout: 258 seconds]
domogled1 is now known as domogled
<gchristensen>
my garage door runs zfs,but it has fairl ylight IO and ram requirements
<superbaloo>
gchristensen: since you're around. where do aarch64 hydra clusters runs?
<superbaloo>
still packet?
<gchristensen>
now known as Equinix Metal, yeah
<superbaloo>
I cant see any arm64 offer since equinix
<gchristensen>
look in AMS1? I just about 10 minutes ago moment ago deprovisioned a c2.large.armthere
<superbaloo>
I just locked at the public website
<gchristensen>
if you still don't see any, you might need to contact support to have them fiddle bits on your account
<gchristensen>
ah okay
<superbaloo>
I havent looked at the admin panel
<superbaloo>
but I was wondering if they planned on keeping them long term
<gchristensen>
pretty sure yes, but that they don't have "third gen" arm hardware yet, so the ydon't advertise it yet
<superbaloo>
I know they invested quite a bit, but :) questions since they just have been acquired :D
<gchristensen>
yeah of cousre
<gchristensen>
I know they're still doing quite a bit with ARM, but we're all reading tea leaves :P
el-hombre-chile has joined #nixos
<superbaloo>
haha
<el-hombre-chile>
hey, I'am having issues trying to build helm 3.5.0 with nix; when it tries to download all of the dependencies I run into this error:
<el-hombre-chile>
github.com/Azure/go-autorest: module github.com/Azure/go-autorest@latest found (v14.2.0+incompatible, replaced by github.com/Azure/go-autorest@v13.3.2+incompatible), but does not contain package github.com/Azure/go-autorest
<sphalerite>
I'm trying to use virtualisation.oci-containers with the podman backend, but my container is failing silently: Jan 15 09:16:23 vivo systemd[1]: podman-home-assistant.service: Main process exited, code=exited, status=125/n/a
m4ts has joined #nixos
<sphalerite>
According to the manpage, 125 means there's an error in podman itself
<sphalerite>
but I find it kind of hard to tell what that is when there's no log output whatsoever besides "Starting podman-home-assistant.service..." "Started podman-home-assistant.service." "podman-home-assistant.service: Main process exited, code=exited, status=125/n/a"
seku has joined #nixos
<sphalerite>
anyone have any debugging suggestions?
cole-h has quit [Ping timeout: 260 seconds]
lsix has joined #nixos
<srk>
strace :)
pjt_tmp has quit [Ping timeout: 264 seconds]
<sphalerite>
srk: how do you trace systemd services?
supersandro2000 has quit [Ping timeout: 265 seconds]
<qyliss>
aforemny: if you want to run exactly what Hydra ran, you can click on "Actions"
<{^_^}>
[nixpkgs] @Mic92 pushed commit from @tpwrules to release-20.09 « lxqt: fix themes and translations »: https://git.io/JtUsu
supersandro2000 has joined #nixos
<aforemny>
qyliss: I tried that and that seems to build nixos/release-combined which succeeds. I have the feeling that does not build the test nixos.tests.networking.scripted.virtual like I want to
<aforemny>
qyliss: Note that `nix-build nixos/tests/networking.nix --arg networkd false -A virtual` builds a different derivation for me on a local checkout of 956afcd. But that succeeds for me as well.
<qyliss>
is your computer particularly fast or slow?
<qyliss>
could be that sort of thing
<aforemny>
qyliss: It might be particular fast!
<aforemny>
I can try on a slower machine.. give me a sec
<qyliss>
the VM tests do a lot of load-bearing sleeps
<Reventlov>
Hey.
civodul has joined #nixos
jmeredith has quit [Quit: Connection closed for inactivity]
oxalica has joined #nixos
<qyliss>
,ask Reventlov
<{^_^}>
Reventlov: Just ask your question. It's the best way to know if anybody can help. Or look at other ways of getting in touch with the Nix community here: https://nixos.wiki/wiki/Get_In_Touch
<Reventlov>
How to copy some private key and certificate you can set up with options to a directory you can also configure via options ? Is there some way to do this without the files to end up in the store ?
tgamblin-llnl has quit [Ping timeout: 264 seconds]
<bigvalen>
Anyone know how https://nixos.wiki/wiki/Installing_from_Linux works ? I'm trying to follow it, and managed to generate my nixos config under /mnt, but running 'nixos-install' complains "error: file 'nixpkgs/nixos' was not found in the Nix search path (add it using $NIX_PATH or -I)"
<bigvalen>
What does it mean by a 'nixpkgs/nixos' *file* ?
<{^_^}>
[nixpkgs] @SuperSandro2000 merged pull request #97101 → python3Packages.wxPython_4_0: build on darwin → https://git.io/JU3Pi
<{^_^}>
[nixpkgs] @SuperSandro2000 pushed 4 commits to master: https://git.io/JtUZF
<qyliss>
bigvalen: what is $NIX_PATH in your environment?
<bigvalen>
qyliss, I'll admit, that's ... one of the more challenging straces I've come across. http://longword.net/~valen/strace.txt - looks like fd 14, which is to /proc/../uid_map. No idea why it's writing to that.
<aforemny>
So, nixos:trunk-combined:tested now only fails on nixos.tests.docker timing out, and that has been fixed in PR #109418 already. So I guess nixos-unstable will succeed at the next evaluation tomorrow? Thanks qyliss for your help and someone who restarted those Hydra jobs. :)
<bigvalen>
Yeah, seems root can't change the uid_map file of anything...even a simple "sleep 200 &" I kicked off. Time to learn how capabilities work..
<ymatsiuk>
I mean the whole idea of review process is eliminated by someone just pushes something directly into master. Don't you think so? You can imagine what would be the impact on project integrity if anyone could just push to master directly:)
tgamblin-llnl has quit [Client Quit]
<supersandro2000>
I would appreciate it if you would at least do a PR
<supersandro2000>
we don't need to do much review but at least check if ofborg has some problem
<simonpe^^>
I have a habit of using buildEnv to "merge" outputs together, but when I do it with some python applications that I created it doesn't work because buildEnv (understandably) erases the shellHook for the derivations so when I try to run the python application in a nix shell its missing all of the enivronment variables so it can't find its dependencies. Is there something else than buildEnv that should be
<exfalso>
Hi, I'm trying to set up nixops to work on a non-NixOS systemd installation. The main piece of work seems to be making systemd units live peacefully next to the "original" units on the OS. I'm currently trying to understand the unit starting/stopping logic in switch-to-configuration.pl, and I have trouble understanding how *fresh*, *new* units are started.
tgamblin-llnl has joined #nixos
<{^_^}>
[nixos-homepage] @github-actions[bot] pushed commit from GitHub Actions to master « Update flake.lock and blogs.xml [ci skip] »: https://git.io/JtUi3
tgamblin-llnl has quit [Client Quit]
<adisbladis>
exfalso: You are trying to deploy to a non-nixos target?
<adisbladis>
If so, that's not supported.
<exfalso>
The main loop seems to iterate over active/activating units that have corresponding /etc/systemd/system/ entries (which fresh units don't), and I don't see any other place where started units are managed
tgamblin-llnl has joined #nixos
<exfalso>
adisbladis: yes, and yes I'm aware
tgamblin-llnl has quit [Client Quit]
hiro99 has joined #nixos
heijligen has joined #nixos
<exfalso>
we kind of need to use non-nixos images because <Big Cloud Provider> does not have SLA on nixos images. I'm trying to avoid simple closure copies+homerolled service logic, it'd be neat if I could simply make nixops work
tgamblin-llnl has joined #nixos
tgamblin-llnl has quit [Client Quit]
<gchristensen>
exfalso: one way to do it would be to copy and install your software and systemd unit files in to a Nix profile, and use https://www.freedesktop.org/software/systemd/man/systemd.generator.html to link them in to the right place in early startup, as if yo uwere managing services with Nix but without replacing the host
<exfalso>
one can get quite far quite quickly, closure copying works as long as nix is installed on the image, but the big unit-diffing logic needs reworking
<{^_^}>
[nixpkgs] @SuperSandro2000 pushed 2 commits to master: https://git.io/JtU90
tgamblin-llnl has joined #nixos
tgamblin-llnl has quit [Client Quit]
tgamblin-llnl has joined #nixos
tgamblin-llnl has quit [Client Quit]
tgamblin-llnl has joined #nixos
tgamblin-llnl has quit [Client Quit]
boxofrox has joined #nixos
tgamblin-llnl has joined #nixos
tgamblin-llnl has quit [Client Quit]
ericsagnes has quit [Ping timeout: 260 seconds]
tgamblin-llnl has joined #nixos
tgamblin-llnl has quit [Client Quit]
tgamblin-llnl has joined #nixos
tgamblin-llnl has quit [Client Quit]
tgamblin-llnl has joined #nixos
tgamblin-llnl has quit [Client Quit]
tgamblin-llnl has joined #nixos
tgamblin-llnl has quit [Client Quit]
tgamblin-llnl has joined #nixos
acarrico has joined #nixos
tgamblin-llnl has quit [Client Quit]
ericsagnes has joined #nixos
tgamblin-llnl has joined #nixos
tgamblin-llnl has quit [Client Quit]
tgamblin-llnl has joined #nixos
tgamblin-llnl has quit [Client Quit]
<tpw_rules>
is there a way to build a shell.nix so i can use it as a package? i have this thing for personal use which has a pile of separate commands and set of python packages. i can do nix-shell thing/shell.nix to get into the environment and do everything but it would be nice to make a command like thing_shell that does the same thing
<bbigras>
Anyone using vault with aws kms to unseal it? When I reboot the vault service fails to start. I think maybe the kms key or whatever if not available right away. If I start the service a couple of minutes later it seems fine.
<tpw_rules>
in particular so i can e.g. include it in environment.systemPackages and make sure it does not get garbage collected
noudle has quit []
tgamblin-llnl has joined #nixos
meh` has joined #nixos
tgamblin-llnl has quit [Client Quit]
<bbigras>
gchristensen: do you use vault with aws kms?
<gchristensen>
no, all local
<gchristensen>
I haven't deployed vault in AWS before, just on bare metal
tgamblin-llnl has joined #nixos
<bbigras>
gchristensen: gotcha. thanks! :)
erasmas has joined #nixos
<gchristensen>
the KMS integration looks great, though
tgamblin-llnl has quit [Client Quit]
<bbigras>
it looks awesome. it unseal automatically. and I guess I could enable the service to restart.
<gchristensen>
yeah, looks good
tgamblin-llnl has joined #nixos
<bbigras>
I'm also considering using sops-nix with KMS. I'm thinking of moving from digital ocean to amazon just for kms.
tgamblin-llnl has quit [Client Quit]
<gchristensen>
KMs and IAM are the real deal
<srk>
what's kms?
<gchristensen>
-llnl is interesting to see here ...
slack1256 has joined #nixos
<srk>
ah, aws kms
<gchristensen>
bbigras: but, not so sure about sops. I think the good part of vault is ephemeral credentials
<bbigras>
key management service. you can use it to decrypt secrets
<srk>
scary :)
<bbigras>
gchristensen: yeah I might just use vault.
<attila_lendvai>
aanderse, thank you for your work on the bee PR! it looks good to my uninitiated eyes, and i've pushed them into the PR. can you please help me with how i can test this on my nixos install? there's plenty of work left, because ethereum's clef needs to be packaged/started with a special setup suitable for bee.
<eyJhb>
Sadly that does not work here supersandro2000
<aanderse>
attila_lendvai: no problem. there are many ways to test nixos modules like `build-vm`, nixos tests, rebuild your system against master, and on
<ronthecookie>
supersandro2000: the 『 character shows my problem off very well
<ronthecookie>
it doesnt seem to happen in my term but it looks bad nearly anywhere else
<attila_lendvai>
aanderse, oh, i see. i cannot just install one service, it brings the entire repo. then i guess i should rebase the changes off of my nixos installs hash (i know how to do this). any hints on how to "rebuild my system against master"? where shall i look?
<aanderse>
attila_lendvai: i'll post what i think is the answer you want in the PR thread for reference
<{^_^}>
[nixpkgs] @SuperSandro2000 pushed 2 commits to master: https://git.io/JtUFy
<{^_^}>
[rfc39-record] @grahamc pushed commit from rfc39 to main « Automated team sync results. »: https://git.io/JtUFQ
veleiro` has joined #nixos
gustavderdrache has quit [Quit: Leaving.]
<arianvp>
Argh; the fact that the --flake argument in nixos-rebuild looks like a flake-ref but isn't one is really confusing. I typed --flake .#nixosConfigurations.t490s but it didn't work and it took my a long time to figure out why not
<arianvp>
why doe we magically expand blah#blah to blah#nixosConfigurations.blah ... this is very confusing to me :/
magnetophon has quit [Ping timeout: 260 seconds]
<cole-h>
arianvp: Yeah, it's a papercut for sure. I think it's just convenience -- if you're building / deploying another system, then `nixosConfigurations.asdf` is a mouthful to type.
<cole-h>
Maybe it should fall back to using the attr if `nixosConfigurations.$attr` errors.
growpotkin has joined #nixos
<arianvp>
Nix is hard to learn because there is so much 'convenience' that makes things look 'easy' (e.g. allowing you to omit config = {} in module systems etc) that I think really hinders my understanding of things
<arianvp>
I finally grasped the whole flake thing; but then I need to remember it's different in this one command :/
<arianvp>
OTOH typing nixosConfigurations every time is also not great
<cole-h>
Which is probably why the convenience of using .#hostname instead of .#nixosConfigurations.hostname was added
<Mic92>
viq|w: you can also check if zfs load-key -a works, because this is what initrd does
<ronthecookie>
supersandro2000: lol, deleting a bunch of fonts fixed it
nij has joined #nixos
<nij>
What's the main difference between Nix and Docker?
<nij>
(as a package manager.)
<cole-h>
I wouldn't consider Docker a package manager in any way, shape, or form, so there's that :P
<viq|w>
Mic92: I was told previously (tested with 20.09) to try adding boot.debug1devices to kernel flags. At which point in that shell, I could see the /dev/vd[a,b] devices, and "zpool import" properly asked for passphrase and continued booting after exiting the shell.
<viq|w>
(never finished booting, IIRC, but that's a separate matter)
<ronthecookie>
nij: that's like comparing apples to oranges
<nij>
ronthecookie: LOL ok I should be more precise.
davidv7 has quit [Remote host closed the connection]
<nij>
I'm trying to introduce friends to Nix.
<nij>
And an obvious selling point is that the system is reproducible.
<attila_lendvai>
aanderse, do you think it's a good idea to use the upstream .service file? it has a line that leads to error: EnvironmentFile=/etc/default/bee
<nij>
But in docker, every dependency is included as well, so reproducibility is achievable too.
<nij>
Why Nix, then?
<viq|w>
nij: because you need some system to run docker *on*
<viq|w>
And that system has configuration.
<ronthecookie>
nij: docker contains everything in containers, is slow AF, makes huge images if you aren't careful and is fairly slow
davidv7 has joined #nixos
<aanderse>
attila_lendvai: i would suggest you submit a PR upstream to replace `EnvironmentFile=/etc/default/bee` with `EnvironmentFile=-/etc/default/bee`
<ronthecookie>
but on nix your binaries are actually all on the same root
<aanderse>
attila_lendvai: to directly answer your question, yes, i think using the upstream systemd unit is a good idea 👍️
<viq|w>
And with nix(os), if you do things correctly, you have whole configuraion in one place. And things that you stop caring about magically stop existing anymore, instead of rotting there, forgotten.
<attila_lendvai>
aanderse, i assume that - sign makes it optional, right?
Cynthia is now known as cynthia_
<nij>
Can I put in this way? Both of them try to achieve reproducibility. Nix does it in a way that allows us to capture the information that we need, in a declarative language. So it is more transparent and thinner.
<nij>
s/need/we really care/
malook has joined #nixos
<aanderse>
attila_lendvai: yes, it means that if the file doesn't exist that is alright
<hiroshi[m]>
nij: nix is a package manager. docker isn't a package manager. there are different things
<viq|w>
I guess nix is more of the docker*file*, though in a declarative way, not imperative like docker is
<viq|w>
well, nixos configuration files.
<viq|w>
Mic92: the link should let you reproduce, if you want to try. I'd love for it to be me doing something wrong, and be told what is it that I'm doing wrong :P
LilleCarl1 has joined #nixos
<nij>
hiroshi[m]: why wouldn't you call docker a package manager?
nij has quit [Remote host closed the connection]
nij has joined #nixos
LilleCarl has quit [Ping timeout: 260 seconds]
<ronthecookie>
wb
<lordcirth>
nij, containers aren't packages?
<hiroshi[m]>
because docker uses an linux images as the base for what you want to build. inside that image you use the package manager of the distro you want. docker creates layers and ns/cgroups for it.
veleiro has joined #nixos
<veleiro>
hello
<lordcirth>
It certainly has similarities - dependency resolution, etc - but calling it a package manager is wierd.
<viq|w>
nij: package managers have no business messing with networking like docker does, for one
<hiroshi[m]>
you can compare nix or guix with rpm or rpm-ostree for example.
<viric>
docker has the same dependency resolution as git commits :)
rb2k has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
davidv7 has quit [Quit: Konversation terminated!]
<hiroshi[m]>
what do you mean with dependency resolution
<viric>
viq|w: docker are cheap VMs where disk images are described by scripts that build them.
<hiroshi[m]>
docker tracks each change that you do from the base image, but that doesn't mean docker is a package manager
<esotericn>
i wouldn't say that docker aims to achieve reproducibility
<hiroshi[m]>
viric: docker aren't cheap vms. contaiers are just cgroups/ns runningin the same base kernel.
<esotericn>
it's possible to use it in such a way, but general practices don't
<viric>
hiroshi[m]: I know. I mean for viq|w to get the picture
<esotericn>
for example, your classic 'from ubuntu:latest apt update'
<hiroshi[m]>
viric: good point. xD
<ronthecookie>
s/apt/run apt/
<viq|w>
It was lordcirth that mentioned dependency resolution regarding docker, so I'm trying to figure out what was meant.
<viric>
ah, "containers aren't packages"? :) containers are the cheap VMs
<nij>
I see.
<viq|w>
Well, I know what containers are, though I don't have much experience with them.
<nij>
"A container image is like a package file, in that it's made up of a other files. But unlike packages, container images don't have a manifest file that contain a name and version. It's essentially a giant zip file that contains all the files an application needs to run, as well as the underlying operating system."
<nij>
So the main difference lies here I think.
<nij>
container v.s. package
<viric>
a container IMAGE. not a container
davidv7 has joined #nixos
<viric>
a container IMAGE is like the disk IMAGE of the VM
<hiroshi[m]>
the main issue with this is that 'docker' isn't one thing. there are so many layers for the name only . hahaha
<nij>
So, like I said, a package is more declarative. It can be described by a manifest. On the other hand, a container is a huge blob.
<viric>
nij: again you mean a container IMAGE
<nij>
viric, my bad. Container Image v.s. Package.
<{^_^}>
[nixpkgs] @ehmry merged pull request #109338 → Incorporate revision in name for Dhall GitHub packages → https://git.io/JtfdP
<tobiasBora>
and after nix-channel --update local-nixos-unstable, then channel points to 257cbbcd3ab according to readlink /nix/var/nix/profiles/per-user/me/channels/local-nixos-unstable
<tobiasBora>
But this commit is 6 days old, while hydra... oups
<tobiasBora>
I understand now
my_nickname has joined #nixos
<tobiasBora>
on hydra the nixos channel is old but not the nixpkgs. My bad
<Reventlov>
"If the .service file is included right in the repository as a file we can directly utilize it in this module."
carthia has quit [Quit: carthia]
ManiacOfMadness has joined #nixos
dunc4n[m] has joined #nixos
mallox has quit [Quit: WeeChat 3.0]
Darkmatter66 has quit [Ping timeout: 256 seconds]
lunik1 has joined #nixos
hiro99 has quit [Ping timeout: 264 seconds]
Fan has quit [Quit: Connection closed]
berberman_ has joined #nixos
berberman has quit [Ping timeout: 260 seconds]
<{^_^}>
[nixpkgs] @domenkozar opened pull request #109469 → move nixos-rebuild into pkgs → https://git.io/JtUpz
<cole-h>
domenkozar[m]: nit: space after lib and before closing curly brace, as well as I think editorconfig will get mad if you don't have an ending newline
<gchristensen>
editorconfig is pretty universal these days
<cole-h>
FWIW I'd rather deal with diffs that include that newline change when someone using editorconfig modifies it than enforce editorconfig in GH Actions. But that's just me.
superherointj has quit [Client Quit]
zupo has quit [Ping timeout: 256 seconds]
lsix1 has quit [Ping timeout: 264 seconds]
lsix1 has joined #nixos
astylian has quit [Remote host closed the connection]
cyphase has quit [Ping timeout: 246 seconds]
cyphase has joined #nixos
<veleiro`>
What are some good nixos-friendly VPS providers in europe?
<cole-h>
domenkozar[m]: Also, secret tip: you can edit your comment and add a space (or whatever, as long as it doesn't change the meaning of the command) to get ofborg to re-parse your comment without needing to post another one
<cole-h>
;)
ymatsiuk has quit [Ping timeout: 248 seconds]
<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « haskell-lukko: jailbreak to fix the build with the latest version of tasty »: https://git.io/JtUjN
<KarlJoad>
cole-h: Oooohh... That is actually super helpful! I didn't know that.
justanotheruser has quit [Ping timeout: 264 seconds]
<{^_^}>
[nixpkgs] @peti pushed to haskell-updates « haskell-http-client-restricted: jailbreak to fix the build with older versions of http-client »: https://git.io/JtTeO
<madmaurice>
Hello, I have a question regarding packaging a program for nixpkgs master. Is this the right channel for that?
<madmaurice>
Or is maybe #nixos-dev?
<colemickens>
madmaurice: yup, go for it
<madmaurice>
Great.
zupo has joined #nixos
ManiacOfMadness has quit [Ping timeout: 272 seconds]
est31 has quit [Ping timeout: 240 seconds]
est31 has joined #nixos
<madmaurice>
I'm trying to package the xed editor from linux mint, it has cinnamon.xapps as dependency but in the latest version it requires a newer version of xapps. Should I to the maintainers or should try to update it myself, in latter case within the cinnamon set or as a new package?
<madmaurice>
Alternatively I could also package an older version of the editor. Would be a couple of months older.
<madmaurice>
s/Should I to/Should I talk to/
<attila_lendvai>
in a nixos service module, is there something like a postInstall hook? alternatively, how can i tell nix that the binary should be available in the preStart environment? pointer to an example is also welcome
Volk has quit [Quit: See you next time!]
OCCASVS has joined #nixos
lsix1 has quit [Ping timeout: 260 seconds]
cole-h has quit [Ping timeout: 240 seconds]
<lordcirth>
madmaurice, if there is a (significantly) newer version of xapps upstream than in master, you should probably talk to the nixos xapps pkg maintainers
<mikroskeem>
pbb: ping, docker stuff blew up again apparently; are you able to check what's up?
<pbb>
mikroskeem: I think that's a user error, but I'll check
<pbb>
mikroskeem: dockerTools builds on master. of course it won't work when you cherry-pick only the PR on top of nixos-unstable, because nixos-unstable doesn't have the Docker 20.10 update
sss2 has joined #nixos
sss2 has quit [Read error: Connection reset by peer]
<{^_^}>
[nixpkgs] @peti merged pull request #109011 → Update Haskell package set to Stackage Nightly 2021-01-11 (plus other fixes) → https://git.io/Jte49
<madmaurice>
lordcirth, It's about 10 versions and half a year behind, so I guess talk to them. What's the best idea here? email or create an issue with mention?
<lordcirth>
madmaurice, issue with @ I think
<madmaurice>
Good, thank you very much!
<{^_^}>
[nixpkgs] @peti pushed 2 commits to haskell-updates: https://git.io/JtTvw
<{^_^}>
[nixpkgs] @peti pushed 2 commits to haskell-updates: https://git.io/JtTvo
<attila_lendvai>
aanderse, please see my question a few lines above. i don't know how to run a script in the service's preStart hook so that the binary is available there
<{^_^}>
[nixpkgs] @peti opened pull request #109474 → Update Haskell package set to Stackage Nightly 2021-01-15 (plus other fixes) → https://git.io/JtTvP
<aanderse>
attila_lendvai: in `preStart` you should reference the package by `${cfg.package}/bin/bee`
astylian has joined #nixos
<attila_lendvai>
aanderse, makes sense. thanks a whole lot!
<aanderse>
no problem, sorry i missed the question first time around :)
<mikroskeem>
pbb: that's what i suspected
<mikroskeem>
alright, all seems good then
davidv7 has quit [Remote host closed the connection]
<attila_lendvai>
aanderse, what can i do if i need some files in my service from an external git repo? (https://github.com/ethersphere/bee-clef/) example packages to look at are also welcome. (this repo only contains some packaging files to run go-ethereum's external signer called clef)
<madmaurice>
What should I use as a version for a package if the package itself does not have any versioning apart from the git revision?
gustavderdrache has joined #nixos
<chrisaw>
madmaurice: I usually use a datestamp in those instances
<chrisaw>
As in DDMMYYYY or similar for the date whatever commit was created
<madmaurice>
Ah that's a good idea. Thanks!
gustavderdrache has quit [Client Quit]
<aanderse>
attila_lendvai: i'm not sure what would be the standard way people would suggest to do that in nixpkgs... given it is from the same developer i would maybe suggest just pulling that service file into the `bee` package itself
<aanderse>
multiple source repos are possible
<{^_^}>
[nixpkgs] @thiagokokada opened pull request #109477 → flutter/dart: remove non-stable versions and bump → https://git.io/JtTJa
ongy[m] has quit [Changing host]
ongy[m] has joined #nixos
ongy[m] has joined #nixos
<attila_lendvai>
aanderse, hrm, yeah, that makes sense, i'll do that. another thing: what is this systemd.packages? do i need to add pkgs.go-ethereum there? or shall i just add it to path = [ cfg.package pkgs.go-ethereum pkgs.coreutils ]; ?
<attila_lendvai>
aanderse, i noticed that if i remove that line, then the service barks that $HOME is not set. it's probably something more fundamental than what i thought
<attila_lendvai>
s/service/bee executable/
gustavderdrache has joined #nixos
thc202 has quit [Quit: thc202]
<aanderse>
attila_lendvai: `systemd.packages` includes any `.service` files (located under $out/lib/systemd/system/) from packages you pass to that option
<aanderse>
so when you removed `cfg.package` from `systemd.packages` in your module, it removed almost all systemd configuration
astylian has quit [Ping timeout: 264 seconds]
erasmas has quit [Quit: leaving]
civodul has quit [Read error: Connection reset by peer]
zarel_ has joined #nixos
zarel has quit [Ping timeout: 265 seconds]
slack1256 has quit [Ping timeout: 246 seconds]
shibboleth has joined #nixos
<supersandro2000>
KarlJoad: 265 commits are a lot
<KarlJoad>
supersandro2000: I know. I have to go through and remove some of them.
hiro98 has quit [Remote host closed the connection]
superherointj has joined #nixos
trubi has joined #nixos
<{^_^}>
[nixpkgs] @fabaff opened pull request #109478 → python3Packages.cbor2: switch to pytestCheckHook → https://git.io/JtTTu
<ronthecookie>
seems not so i guess this is the best place to ask
<cemguresci[m]>
:D
superherointj_ has quit [Quit: Leaving]
<attila_lendvai>
aanderse, in bee's postInstall, why didn't you just delete the entire ExecStart line? AFAIU, it's overridden anyway in overrides.conf, no?
eganjs[m] has joined #nixos
<ronthecookie>
anyone know where to get in touch with the nixos infra people?
<kini>
>If a package is not a release but a commit from a repository, then the version part of the name must be the date of that (fetched) commit. The date must be in "YYYY-MM-DD" format. Also append "unstable" to the name - e.g., "pkgname-unstable-2014-09-23".
seku has quit [Quit: Connection closed]
astylian has joined #nixos
<madmaurice>
@kini I see. Thanks!
vidbina_ has quit [Ping timeout: 264 seconds]
mmmattyx has quit [Quit: Connection closed for inactivity]
diamondman has quit [Read error: Connection reset by peer]
mudri has quit [Read error: Connection reset by peer]
gausby has quit [Read error: Connection reset by peer]
yrashk has quit [Read error: Connection reset by peer]
fnords has quit [Write error: Connection reset by peer]
cynthia_ has quit [Write error: Connection reset by peer]
dsal has quit [Read error: Connection reset by peer]
fnords has joined #nixos
gausby has joined #nixos
pikajude has joined #nixos
pingiun has joined #nixos
diamondman has joined #nixos
dsal has joined #nixos
shibboleth has quit [Quit: shibboleth]
cynthia_ has joined #nixos
mudri has joined #nixos
yrashk has joined #nixos
lorenzleutgeb[m] has quit [Ping timeout: 246 seconds]
dbirks[m] has quit [Ping timeout: 246 seconds]
Kimat[m] has quit [Ping timeout: 246 seconds]
Fafnir[m] has quit [Ping timeout: 246 seconds]
scheibenkleister has quit [Ping timeout: 246 seconds]
felschr[m] has quit [Ping timeout: 246 seconds]
dxb[m] has quit [Ping timeout: 246 seconds]
tsmeets[m] has quit [Ping timeout: 246 seconds]
solomon[m] has quit [Ping timeout: 246 seconds]
boogiewoogie[m] has quit [Ping timeout: 246 seconds]
fc[m]1 has quit [Ping timeout: 246 seconds]
Sajesajama[m] has quit [Ping timeout: 246 seconds]
eadwu has quit [Ping timeout: 246 seconds]
alexrelis[m] has quit [Ping timeout: 246 seconds]
deselby has quit [Ping timeout: 246 seconds]
mhj[m] has quit [Ping timeout: 246 seconds]
neuronuser1[m] has quit [Ping timeout: 246 seconds]
coolneng[m] has quit [Ping timeout: 246 seconds]
botayhard[m] has quit [Ping timeout: 246 seconds]
OJ[m] has quit [Ping timeout: 246 seconds]
M3nternamehere[m has quit [Ping timeout: 246 seconds]
mickbrooks[m] has quit [Ping timeout: 246 seconds]
rhyfore[m] has quit [Ping timeout: 246 seconds]
nickbublik[m] has quit [Ping timeout: 246 seconds]
ptitfred[m] has quit [Ping timeout: 246 seconds]
totoroot[m] has quit [Ping timeout: 246 seconds]
guixos[m] has quit [Ping timeout: 246 seconds]
psibi[m] has quit [Ping timeout: 246 seconds]
fadenb[m] has quit [Ping timeout: 246 seconds]
kateskips has quit [Ping timeout: 246 seconds]
M86ul[m]1 has quit [Ping timeout: 246 seconds]
shu9 has quit [Ping timeout: 246 seconds]
ongy[m] has quit [Ping timeout: 246 seconds]
musicmatze has quit [Ping timeout: 246 seconds]
siraben has quit [Ping timeout: 246 seconds]
daemon1024 has quit [Ping timeout: 246 seconds]
dtz has quit [Ping timeout: 246 seconds]
danielrf[m] has quit [Ping timeout: 246 seconds]
leibniz[m] has quit [Ping timeout: 246 seconds]
rummik has quit [Ping timeout: 246 seconds]
<madmaurice>
O_o
cemguresci[m] has quit [Ping timeout: 246 seconds]
tuxcoder[m] has quit [Ping timeout: 246 seconds]
faya01[m] has quit [Ping timeout: 246 seconds]
deoxal[m] has quit [Ping timeout: 246 seconds]
ms[m] has quit [Ping timeout: 246 seconds]
Bla[m] has quit [Ping timeout: 246 seconds]
puzzlewolf has quit [Ping timeout: 246 seconds]
Poscat[m] has quit [Ping timeout: 246 seconds]
Ke has quit [Ping timeout: 246 seconds]
mikroskeem has quit [Ping timeout: 246 seconds]
avalenn has quit [Ping timeout: 246 seconds]
kraem has quit [Ping timeout: 246 seconds]
<madmaurice>
O.o
mikroskeem has joined #nixos
slack1256 has joined #nixos
jess has quit [Quit: Leaving]
<supersandro2000>
kini: that part should be really be more specific
<supersandro2000>
it mentions nothing about pname version
<samueldr>
written before pname was a thing
Kimat[m] has joined #nixos
deselby has joined #nixos
lorenzleutgeb[m] has joined #nixos
Sajesajama[m] has joined #nixos
dbirks[m] has joined #nixos
Fafnir[m] has joined #nixos
<attila_lendvai>
aanderse, do you have any idea why i'm getting this? https://pastebin.com/7ws3vpTg if i keep the tmp directory, then it weirdly contains an extra zero length file: sysroots.tar.bz2
scheibenkleister has joined #nixos
fc[m]1 has joined #nixos
eadwu has joined #nixos
felschr[m] has joined #nixos
boogiewoogie[m] has joined #nixos
tsmeets[m] has joined #nixos
solomon[m] has joined #nixos
OJ[m] has joined #nixos
M3nternamehere[m has joined #nixos
rhyfore[m] has joined #nixos
nickbublik[m] has joined #nixos
ptitfred[m] has joined #nixos
dxb[m] has joined #nixos
alexrelis[m] has joined #nixos
rummik has joined #nixos
botayhard[m] has joined #nixos
psibi[m] has joined #nixos
ongy[m] has joined #nixos
kateskips has joined #nixos
M86ul[m]1 has joined #nixos
mickbrooks[m] has joined #nixos
<attila_lendvai>
i have no idea how to debug this
daemon1024 has joined #nixos
guixos[m] has joined #nixos
totoroot[m] has joined #nixos
mhj[m] has joined #nixos
neuronuser1[m] has joined #nixos
coolneng[m] has joined #nixos
musicmatze has joined #nixos
shu9 has joined #nixos
karantan has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
danielrf[m] has joined #nixos
fadenb[m] has joined #nixos
<{^_^}>
[nixpkgs] @fabaff opened pull request #109482 → python3Packages.aiocoap: init at 0.4b3 → https://git.io/JtTmT
dtz has joined #nixos
siraben has joined #nixos
leibniz[m] has joined #nixos
qz has quit [Quit: WeeChat 2.9]
faya01[m] has joined #nixos
deoxal[m] has joined #nixos
gluegadget has quit [Ping timeout: 268 seconds]
rpgwaiter has quit [Ping timeout: 268 seconds]
kristjansson has quit [Ping timeout: 268 seconds]
pointfree has quit [Ping timeout: 268 seconds]
manveru has quit [Ping timeout: 268 seconds]
tdeo has quit [Read error: Connection reset by peer]
hax404 has quit [Remote host closed the connection]
M86ul[m]1 has quit [Ping timeout: 268 seconds]
iuiiik[m] has quit [Ping timeout: 268 seconds]
mupf[m] has quit [Ping timeout: 268 seconds]
arghzero[m] has quit [Ping timeout: 268 seconds]
stu_[m] has quit [Ping timeout: 268 seconds]
xd-inc[m] has quit [Ping timeout: 268 seconds]
davegallant[m] has quit [Ping timeout: 268 seconds]
tgys[m] has quit [Ping timeout: 268 seconds]
msciabarra[m] has quit [Ping timeout: 268 seconds]
regnat has quit [Ping timeout: 268 seconds]
us3r[m] has quit [Ping timeout: 268 seconds]
l33[m] has quit [Ping timeout: 268 seconds]
rednaZ[m] has quit [Ping timeout: 268 seconds]
emily has quit [Ping timeout: 268 seconds]
hiroshi[m] has quit [Ping timeout: 268 seconds]
jasom has quit [Ping timeout: 268 seconds]
siel has quit [Ping timeout: 268 seconds]
globin has quit [Ping timeout: 268 seconds]
Reventlov has quit [Ping timeout: 268 seconds]
benley has quit [Ping timeout: 268 seconds]
globin has joined #nixos
lucus16 has joined #nixos
samhza has quit [Read error: Connection reset by peer]
alunduil has quit [Ping timeout: 260 seconds]
ProofTechnique has quit [Ping timeout: 260 seconds]
samhza has joined #nixos
siel has joined #nixos
Dagger has joined #nixos
mauli_ has joined #nixos
wildsebastian has quit [Read error: Connection reset by peer]
pingveno has quit [Read error: Connection reset by peer]
elvishjerricco has quit [Read error: Connection reset by peer]
lstanley has quit [Read error: Connection reset by peer]
mankyKitty has quit [Read error: Connection reset by peer]
carter has quit [Read error: Connection reset by peer]
rizary has quit [Read error: Connection reset by peer]
kitemikaze_ has quit [Read error: Connection reset by peer]
aria has quit [Read error: Connection reset by peer]
aria has joined #nixos
ProofTechnique has joined #nixos
pingveno has joined #nixos
elvishjerricco has joined #nixos
wildsebastian has joined #nixos
mankyKitty has joined #nixos
rizary has joined #nixos
morr has joined #nixos
carter has joined #nixos
lstanley has joined #nixos
fendor has quit [Remote host closed the connection]
kristjansson has joined #nixos
alunduil has joined #nixos
kristjansson has quit [Excess Flood]
kitemikaze_ has joined #nixos
rot13 has quit [Ping timeout: 268 seconds]
lopsided98 has quit [Remote host closed the connection]
beaky has joined #nixos
Heirlung has joined #nixos
aswanson has quit [Ping timeout: 256 seconds]
neuronuser1[m] has quit [Ping timeout: 260 seconds]
kateskips has quit [Ping timeout: 260 seconds]
daemon1024 has quit [Ping timeout: 260 seconds]
psibi[m] has quit [Ping timeout: 260 seconds]
felschr[m] has quit [Ping timeout: 260 seconds]
boogiewoogie[m] has quit [Ping timeout: 260 seconds]
dxb[m] has quit [Ping timeout: 260 seconds]
botayhard[m] has quit [Ping timeout: 260 seconds]
pbr_1973[m] has quit [Ping timeout: 260 seconds]
eganjs[m] has quit [Ping timeout: 260 seconds]
Kimat[m] has quit [Ping timeout: 260 seconds]
anonymouserobot[ has quit [Ping timeout: 260 seconds]
grahamc[m] has quit [Ping timeout: 260 seconds]
spinlock[m] has quit [Ping timeout: 260 seconds]
qp12[m] has quit [Ping timeout: 260 seconds]
jakosimov[m] has quit [Ping timeout: 260 seconds]
AlesHuzik[m] has quit [Ping timeout: 260 seconds]
fufexan[m] has quit [Ping timeout: 260 seconds]
lurpahi has quit [Ping timeout: 260 seconds]
Zoom[m] has quit [Ping timeout: 260 seconds]
lostnet[m] has quit [Ping timeout: 260 seconds]
Lorenzoas3ii[m] has quit [Ping timeout: 260 seconds]
evhan[m] has quit [Ping timeout: 260 seconds]
kfr[m] has quit [Ping timeout: 260 seconds]
kevincox[m] has quit [Ping timeout: 260 seconds]
codebam has quit [Ping timeout: 260 seconds]
nikola[m] has quit [Ping timeout: 260 seconds]
nilsirl[m] has quit [Ping timeout: 260 seconds]
leonardp has quit [Ping timeout: 260 seconds]
tcakoi has quit [Ping timeout: 260 seconds]
nikivi[m] has quit [Ping timeout: 260 seconds]
aveeran[m] has quit [Ping timeout: 260 seconds]
timokau[m] has quit [Ping timeout: 260 seconds]
garbas[m] has quit [Ping timeout: 260 seconds]
davidak[m] has quit [Ping timeout: 260 seconds]
domenkozar[m] has quit [Ping timeout: 260 seconds]
maralorn has quit [Ping timeout: 260 seconds]
chreekat[m] has quit [Ping timeout: 260 seconds]
faya01[m] has quit [Ping timeout: 260 seconds]
fadenb[m] has quit [Ping timeout: 260 seconds]
coolneng[m] has quit [Ping timeout: 260 seconds]
ptitfred[m] has quit [Ping timeout: 260 seconds]
rhyfore[m] has quit [Ping timeout: 260 seconds]
M3nternamehere[m has quit [Ping timeout: 260 seconds]
helithumper[m] has quit [Ping timeout: 260 seconds]
edude03 has quit [Ping timeout: 260 seconds]
lycium[m] has quit [Ping timeout: 260 seconds]
ziguana[m] has quit [Ping timeout: 260 seconds]
noneucat has quit [Ping timeout: 260 seconds]
Dandellion has quit [Ping timeout: 260 seconds]
hsiktas[m] has quit [Ping timeout: 260 seconds]
Notkea[m] has quit [Ping timeout: 260 seconds]
lopsided98 has joined #nixos
<aanderse>
attila_lendvai: paste expired
<abathur>
supersandro2000: stumbled on rs-git-fsmonitor today; curious if you're handling the git-config fraction of it programmatically, or just manually? (I may be interested in consuming it as a dependency for another package, but I'm not sure there's a sane Nixy way to handle the config)
<supersandro2000>
at which point you could just tell them what you want in the config
turlando has quit [Read error: Connection reset by peer]
yurb has joined #nixos
sefu[m] has joined #nixos
turlando has joined #nixos
<attila_lendvai>
aanderse, ok, i'll push what i have into the PR and leave a comment
<aanderse>
attila_lendvai: thanks
<abathur>
supersandro2000: nod, I've been poking at a statusd alternative which can beat it in some cases and am curious to see if rs-git-fsmonitor improves performance on the cases where it doesn't
<supersandro2000>
I don't think it is faster
<supersandro2000>
watchman can also use quite some file descriptors
linus0[m] has joined #nixos
<supersandro2000>
doing it globally broke VSCode for me very often
<abathur>
hmm, yeah
<supersandro2000>
there is also a perl hook which achieves something similar
<attila_lendvai>
aanderse, oooh, could it be that mkDerivation picks up on a Makefile and calls it in a DWIM'ish way?
Brio[m] has joined #nixos
<supersandro2000>
gitstatusd does not track all feature of git status and sometimes yields strange things compared to git status
<attila_lendvai>
aanderse, yeah, that's the issue. now, how can i tell it not to do that...?
mica[m] has joined #nixos
revity[m] has joined #nixos
Yakulu[m] has joined #nixos
Sajesajama[m] has joined #nixos
Ollie[m]1 has joined #nixos
stu_[m] has joined #nixos
msciabarra[m] has joined #nixos
arghzero[m] has joined #nixos
lambdanature[m] has joined #nixos
us3r[m] has joined #nixos
tgys[m] has joined #nixos
mupf[m] has joined #nixos
Guest98258 has quit [*.net *.split]
ktosiek has quit [*.net *.split]
Cale has quit [*.net *.split]
das_j has quit [*.net *.split]
sjlnk_ has quit [*.net *.split]
hauleth has quit [*.net *.split]
cruxeter has quit [*.net *.split]
Maxdamantus has quit [*.net *.split]
koluacik has quit [*.net *.split]
MarcWebe1 has quit [*.net *.split]
energizer has quit [*.net *.split]
pbogdan has quit [*.net *.split]
sknebel has quit [*.net *.split]
LangeOortjes has quit [*.net *.split]
addcninb- has quit [*.net *.split]
sumner has quit [*.net *.split]
patagonicus has quit [*.net *.split]
brettgilio has quit [*.net *.split]
Gohla has quit [*.net *.split]
_rvl has quit [*.net *.split]
m1cr0man has quit [*.net *.split]
concept2 has quit [*.net *.split]
ornxka has quit [*.net *.split]
ivan has quit [*.net *.split]
adamCS has quit [*.net *.split]
orbekk has quit [*.net *.split]
_habnabit has quit [*.net *.split]
edef has quit [*.net *.split]
xdej has quit [*.net *.split]
bob_twinkles has quit [*.net *.split]
maurer has quit [*.net *.split]
xd-inc[m] has joined #nixos
doublej41 has quit [*.net *.split]
mla has quit [*.net *.split]
cruxeternus has joined #nixos
edef has joined #nixos
sjlnk has joined #nixos
_habnabit has joined #nixos
ivan has joined #nixos
bob_twinkles has joined #nixos
davidv7 has quit [Remote host closed the connection]
sknebel has joined #nixos
clever has joined #nixos
xdej_ has joined #nixos
clever has quit [Changing host]
MarcWeber has joined #nixos
clever has joined #nixos
<abathur>
basically I can beat gitstatusd by a lot in a few cases when I don't need to fall back on running `git diff --quiet...` (in which case it performs somewhere between gitstatus and a plain `git status | parse` approach)
l33[m] has joined #nixos
Maxdamantus has joined #nixos
maurer has joined #nixos
addcninb- has joined #nixos
Cale has joined #nixos
patagonicus has joined #nixos
sumner has joined #nixos
hauleth has joined #nixos
concept2 has joined #nixos
ktosiek has joined #nixos
ornxka has joined #nixos
davegallant[m] has joined #nixos
regnat has joined #nixos
Guest98258 has joined #nixos
m1cr0man has joined #nixos
qqlq has joined #nixos
orbekk has joined #nixos
Vanilla[m] has joined #nixos
das_j has joined #nixos
doublej41 has joined #nixos
slby[m] has joined #nixos
b1nch0[m] has joined #nixos
koluacik has joined #nixos
<abathur>
so I'm curious about the prospect that fsmonitor could speed up those diff cases if it knows there aren't changes
<supersandro2000>
and this are the only two examples I could find
Kimat[m] has joined #nixos
Lorenzoas3ii[m] has joined #nixos
tcakoi has joined #nixos
felschr[m] has joined #nixos
boogiewoogie[m] has joined #nixos
kateskips has joined #nixos
hspak4 has joined #nixos
neuronuser1[m] has joined #nixos
<abathur>
supersandro2000: detect the cases where I can beat it, or where I can't? :)
daemon1024 has joined #nixos
evhan[m] has joined #nixos
AlesHuzik[m] has joined #nixos
Zoom[m] has joined #nixos
jakosimov[m] has joined #nixos
lurpahi has joined #nixos
timokau[m] has joined #nixos
lostnet[m] has joined #nixos
kevincox[m] has joined #nixos
anonymouserobot[ has joined #nixos
grahamc[m] has joined #nixos
maralorn has joined #nixos
helithumper[m] has joined #nixos
edude03 has joined #nixos
ziguana[m] has joined #nixos
noneucat has joined #nixos
rhyfore[m] has joined #nixos
domenkozar[m] has joined #nixos
M3nternamehere[m has joined #nixos
Dandellion has joined #nixos
aveeran[m] has joined #nixos
fadenb[m] has joined #nixos
hspak has quit [Ping timeout: 246 seconds]
hspak4 is now known as hspak
fufexan[m] has joined #nixos
coolneng[m] has joined #nixos
faya01[m] has joined #nixos
<abathur>
here's a little recording I made from the initial draft of it https://twitter.com/abathur/status/1342687137242734594 (I am working on tests and actual benchmarking now, and have identified some things that need fixing)
ptitfred[m] has joined #nixos
lycium[m] has joined #nixos
Notkea[m] has joined #nixos
hsiktas[m] has joined #nixos
<abathur>
I see ironing those out as a next step before I try to promote it a little and fish for some help
Volk has joined #nixos
lsix1 has quit [Ping timeout: 260 seconds]
AWizzArd has quit [Changing host]
AWizzArd has joined #nixos
lsix1 has joined #nixos
astylian has quit [Ping timeout: 240 seconds]
gustavderdrache has joined #nixos
<abathur>
hmm, some naive manual testing suggests the things I might run are as slow or slower with the fsmonitor; I guess it may be more of a benefit for spinning metal
Guest98258 has quit [Changing host]
Guest98258 has joined #nixos
Guest98258 is now known as Reventlov
lord| has joined #nixos
Dotz0cat has joined #nixos
Rusty1 has joined #nixos
<abathur>
ah!
<abathur>
I'm using git 2.29.2, but I did find a patchset supposedly in git 2.30.0 that addresses fsmonitor data not being used in git diff...